providers play in deciding whether to treat an infant who is born alive but in distress and is so premature that, despite advancements in neonatal intensive care, [he or she] has a largely uncertain prognosis” [16]. The court determined that “... a physician who is confronted with emergent circumstances and provides life-sustaining treatment to a mi- 72 S. Van McCrary / Neonatology and the specter of the law nor child, is not liable for first obtaining consent from the parents” [16]. In other words, informed consent for resuscitation is not required in such emergent circumstances because the parents do not have the option of refusing. The decision is a narrow one and provides little guidance except that neonatologists are permitted to make a decision about resuscitation immediately after birth in cases of extreme prematurity. As George Annas has observed, “... nothing in the decision requires the presence of a neonatologist at the delivery. More troubling, the court implies that life is always preferable to death for a newborn and could be interpreted in the future to support the neonatologist who always resuscitates newborns, no matter how premature or how unlikely their survival is without severe disabilities” [16]. It is notable that both appellate court decisions have interpreted the applicability of the Baby Doe rules to cases of extreme preterm birth rather than the moderate genetic anomalies that first spurred the passage of the rules. Because there is little significant opposition today (among the public or medical community) to providing full treatment for infants with moderate health problems like those in the cases that originally generated the rules, while cases of extreme prematurity present risks of catastrophic and permanent disability, it may be time to consider whether the same criteria should be applied or whether humane public policy requires a revised set of criteria. Kopelman concludes her arguments with a statement of principles that clearly describes the goal: “The Baby Doe rules should be challenged by the AAP because they impede individualized and compassionate care for children advocated by the AAP, and they give too little consideration to parental consent, clinical judgment, and duties to minimize unnecessary suffering and treat others the way we wish to be treated” [17]. 3. Cultures of neonatology and the prediction of clinical outcomes A central factor in the social policy struggle about laws regulating the practice of neonatology is the uncertainty among physicians when attempting to predict morbidity and mortality among infants born with extreme prematurity. This controversy is further complicated by high variability in the outcomes reported in the medical literature from developed countries, especially at the threshold of viability – births at < 23 weeks gestational age. For Europe, data from the EPICure Study Group in the United Kingdom and Ireland (published in 2000 and 2005), is currently thought to be the most valid and reliable in this population [19–21]. Among 22 week gestation infants born alive, the EPICure data indicate a prevalence of survival-to-discharge of one percent and survival without overall disability at 30 months at five percent of NICU admissions (0.7 percent of live births) [20]. Follow-up data from EPICure at age six years indicates a prevalence of survival without overall disability at 12 percent within this tiny population (0.0008 percent of live births) [21]. In stark contrast to the EPICure results, are data on 22 week premature births from Japan (collected only from government-recognized perinatal centers) suggesting an overall 28 day survival rate of approximately 32 percent [22]. Other research, from the VermontOxford Network, has reported survival to discharge only in terms of birth weight rather than gestational age; this group found infants born at < 500 grams (50th percentile at 22 weeks) too immature to survive with current technology [23]. Another group, Seri and Evans, has combined these criteria to constitute what they call a “gray zone” of decision making, and proposed a dual measure of < 23 weeks gestation and < 500 grams [24]. These varying methodologies have highlighted the depth of lingering uncertainty about the best measures for evaluating this population. Further complicating matters are anecdotal reports from professional conferences suggesting that in some parts of Europe, infants born at < 23 weeks are routinely treated with good results (personal communication from Stephen M. Baumgart, MD). However difficult it is to predict survival, predicting severe morbidity appears less so. John D. Lantos has noted that there is a group of about 10 percent of neonates about whom predicting survival is “little better than chance” [25]. Yet, when it comes to predicting severe disability the predictions are considerably more reliable. Rates of survival (among all live births) with no disability at six years in the EPICure study population were 0.0008 percent at 22 weeks gestation,one percent at 23 weeks, and three percent at 24 weeks – data that Vohr and Allen have described as “troubling” [26]. Lantos and colleagues have concluded that “... clinical predictors of medical futility are not very accurate if futility refers only to survival. However, if it refers to neurologically intact survival, the doctors and nurses were very good at foreseeing the future” [25,27]. This phenomenon makes